Added PremulOutput and UnpremulOutput FP helpers. These are used
(rather than GrConfigConversionEffect) when working with FP16
textures (and will also be used for other configs that can't be
round-tripped via rounding).
BUG=skia:5853
Change-Id: I101592c26c4f0b379d5e5a8678ef7b2f08e6ad56
Reviewed-on: https://skia-review.googlesource.com/9980
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
This silences a new warning in clang 5.0
Change-Id: Ieb5b75a6ffed60107c3fd16075d2ecfd515b55e8
Reviewed-on: https://skia-review.googlesource.com/10006
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Mike Klein <mtklein@chromium.org>
This reverts commit fbb56ce83f.
Reason for revert: Chrome's ancient libstdc++ doesn't like use of steady_clock
Original change's description:
> Add a new GrResourceCache purging mechanism for purging unused resources.
>
> The client may call GrContext::purgeResourceNotUsedSince() with a stead_clock::time_point and all resources that have been purgeable since before that time point are purged.
>
> This is intended to replace the "max unused flushes" purging mechanism once Chrome adopts it.
>
> Change-Id: I28881dd2959cc01c0acca81b2d6001ee5626439d
> Reviewed-on: https://skia-review.googlesource.com/8920
> Commit-Queue: Brian Salomon <bsalomon@google.com>
> Reviewed-by: Robert Phillips <robertphillips@google.com>
> Reviewed-by: Eric Karl <ericrk@google.com>
>
TBR=bsalomon@google.com,robertphillips@google.com,ericrk@google.com,reviews@skia.org,ericrk@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Change-Id: I3baa3b2e80302315b757d4d625732459e13795d4
Reviewed-on: https://skia-review.googlesource.com/10004
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
The client may call GrContext::purgeResourceNotUsedSince() with a stead_clock::time_point and all resources that have been purgeable since before that time point are purged.
This is intended to replace the "max unused flushes" purging mechanism once Chrome adopts it.
Change-Id: I28881dd2959cc01c0acca81b2d6001ee5626439d
Reviewed-on: https://skia-review.googlesource.com/8920
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
Reviewed-by: Eric Karl <ericrk@google.com>
* Translate to python, use multiprocessing to make it wun in reasonanble time.
* Fix three headers.
* Move one header.
Change-Id: I0b26f912bb8086ec158d1c0bae76b923ec6121a4
Reviewed-on: https://skia-review.googlesource.com/8490
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Hal Canary <halcanary@google.com>
This also subtlely allows clients to convert between F32 and F16.
BUG=skia:
CQ_INCLUDE_TRYBOTS=skia.primary:Test-Ubuntu-GCC-GCE-CPU-AVX2-x86_64-Release-SKNX_NO_SIMD
Change-Id: Ied5f2295fce00c69d8cf85730be899f3f8597915
Reviewed-on: https://skia-review.googlesource.com/9914
Reviewed-by: Mike Reed <reed@google.com>
Reviewed-by: Mike Klein <mtklein@chromium.org>
Commit-Queue: Matt Sarett <msarett@google.com>
This CL removes the GrTexture-based ctor forcing everyone to create deferred SkImage_Gpus.
split out into:
https://skia-review.googlesource.com/c/9106/ (Remove atlas creation from GrResourceProvider)
Change-Id: I266bbe089c242fe54d5b7adcc7895aa5a39440a0
Reviewed-on: https://skia-review.googlesource.com/6680
Commit-Queue: Robert Phillips <robertphillips@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
We don't support unpremultiplied images. Therefore:
- Don't test unpremul source images.
- After doing an unpremultiplying read, make sure to premul before drawing
For this to work with F16, add support for F16 sources to
SkColorSpaceXform.
Public API change is comments-only.
TBR=reed@google.com
BUG=skia:
Change-Id: Ie05b58231e99ca88cd7792b65ffbb4f390b01726
Reviewed-on: https://skia-review.googlesource.com/9900
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Matt Sarett <msarett@google.com>
This will be rolled out in three stages:
1) make SkMemory.h and have SkTypes.h include it.
2) Adjust chromium and android.
3) no long include SkMemory.h in SkTypes.h
Change-Id: If360ef5e1164d88f50b03f279e2e963ca2f57d5d
Reviewed-on: https://skia-review.googlesource.com/9874
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Herb Derby <herb@google.com>
Slight variation on https://skia-review.googlesource.com/c/7346/,
which had to be reverted. This version also changes the non-mip
upload path to use a GrTextureProxy. Hopefully this doesn't trigger
the same perf regression that we saw last time.
BUG=skia:
Change-Id: Ic99bdd93904089942fb680c3af9c4c36310074c0
Reviewed-on: https://skia-review.googlesource.com/9892
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Osman <brianosman@google.com>
Skia needs implementations (ports) of the declarations in this header,
but no one outside of Skia the library should use them.
Change-Id: I1b92eca787a61892d7b3c0ac8c5996774429ae24
Reviewed-on: https://skia-review.googlesource.com/9878
Commit-Queue: Ben Wagner <bungeman@google.com>
Reviewed-by: Mike Reed <reed@google.com>
Reviewed-by: Leon Scroggins <scroggo@google.com>
This unused feature complicates MDB.
Chrome compiles locally for me with this CL.
frameworks/base/libs/hwui compiles locally for me with this CL.
Change-Id: Id3ad64dac72eace52855896df0d7ce3679f15884
Reviewed-on: https://skia-review.googlesource.com/9882
Reviewed-by: Brian Salomon <bsalomon@google.com>
Reviewed-by: Matt Sarett <msarett@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
The fix was to release the array of vertices in the picturerecorder
destructor (where we also release textblobs etc.
This reverts commit 1eb3fef136.
BUG=skia:
Change-Id: I3bf4acd6ad209205b0832a3cb7f94cd89dfcefc5
Reviewed-on: https://skia-review.googlesource.com/9826
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Reed <reed@google.com>
This reverts commit 14583e11fd.
Reason for revert: leaking
Direct leak of 499104 byte(s) in 2112 object(s) allocated from:
#0 0x1e195f0 in operator new(unsigned long) (/b/swarm_slave/w/irazbR79/out/Debug/dm+0x1e195f0)
#1 0x3142b0a in SkVertices::Builder::init(SkCanvas::VertexMode, int, int, SkVertices::Sizes const&) (/b/swarm_slave/w/irazbR79/out/Debug/dm+0x3142b0a)
Original change's description:
> More SkVertices implementation work
>
> - change virtuals to take const SkVertices*, as we do for TextBobs and Images
> - override onDrawVerticesObject in recording canvases
> - deserialize raw-vertices into SkVertices object
>
> Possibly a follow-on would intercept the raw-form directly in canvas,
> and remove the virtual, and only support the object form.
>
> BUG=skia:6366
>
> Change-Id: I57a932667ccb3b3b004beb802ac3ae6898e3c6e0
> Reviewed-on: https://skia-review.googlesource.com/9633
> Commit-Queue: Mike Reed <reed@google.com>
> Reviewed-by: Brian Salomon <bsalomon@google.com>
>
TBR=bsalomon@google.com,reed@google.com,reviews@skia.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:6366
Change-Id: I40bb7a20698ef6aa0a9ef71a3d6ac4c1473e081c
Reviewed-on: https://skia-review.googlesource.com/9825
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Reed <reed@google.com>
- change virtuals to take const SkVertices*, as we do for TextBobs and Images
- override onDrawVerticesObject in recording canvases
- deserialize raw-vertices into SkVertices object
Possibly a follow-on would intercept the raw-form directly in canvas,
and remove the virtual, and only support the object form.
BUG=skia:6366
Change-Id: I57a932667ccb3b3b004beb802ac3ae6898e3c6e0
Reviewed-on: https://skia-review.googlesource.com/9633
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
Deferring removing the flags parameter from the virtuals until a
later CL (as it collides with another related CL)
BUG=skia:6366
Change-Id: I817fae3df03ecebe5ec3532f691ed06deab890e6
Reviewed-on: https://skia-review.googlesource.com/9739
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Mike Reed <reed@google.com>
This reverts commit 2bf4b3a97b.
Reason for revert: nanobench assertion
Original change's description:
> Remove GrPipeline from GrDrawOp.
>
> GrDrawOp subclasses are now free to construct their pipelines at flush time and now in theory could use multiple GrPipelines for multipass rendering.
>
> GrProcessorSet may be used to retain the processors from a GrPaint with "pending execution" style refs.
>
> NVPR and Instanced rendering are updated to create their pipelines at flush time without a GrPipelineBuilder.
>
> The monolithic pipeline creation/management that was on GrDrawOp is moved to GrMeshDrawOp. However, this is temporary and will be removed in coming changes.
>
> Change-Id: I124282e3cea5d070970b5460c8a679fcaf7a8eff
> Reviewed-on: https://skia-review.googlesource.com/7279
> Commit-Queue: Brian Salomon <bsalomon@google.com>
> Reviewed-by: Robert Phillips <robertphillips@google.com>
>
TBR=bsalomon@google.com,robertphillips@google.com,csmartdalton@google.com,reviews@skia.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Change-Id: I1bc64f6cbbd5f482417637a034342c2b5371dc5c
Reviewed-on: https://skia-review.googlesource.com/9817
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
GrDrawOp subclasses are now free to construct their pipelines at flush time and now in theory could use multiple GrPipelines for multipass rendering.
GrProcessorSet may be used to retain the processors from a GrPaint with "pending execution" style refs.
NVPR and Instanced rendering are updated to create their pipelines at flush time without a GrPipelineBuilder.
The monolithic pipeline creation/management that was on GrDrawOp is moved to GrMeshDrawOp. However, this is temporary and will be removed in coming changes.
Change-Id: I124282e3cea5d070970b5460c8a679fcaf7a8eff
Reviewed-on: https://skia-review.googlesource.com/7279
Commit-Queue: Brian Salomon <bsalomon@google.com>
Reviewed-by: Robert Phillips <robertphillips@google.com>
This reverts commit b64bcbdc3a.
Reason for revert:
Android build failed as shown below.
frameworks/base/libs/hwui/VkLayer.cpp:32:41: error: too many arguments to function call, expected 0, have 1
mImage = surface->makeImageSnapshot(SkBudgeted::kNo);
Original change's description:
> Remove budgeted parameter from SkSurface::makeImageSnapshot
>
> This unused feature complicates MDB.
>
> Chrome compiles locally for me with this CL.
>
> Change-Id: I611e464885fb984030eace43ead42cf39d0e7f72
> Reviewed-on: https://skia-review.googlesource.com/9734
> Reviewed-by: Brian Salomon <bsalomon@google.com>
> Commit-Queue: Robert Phillips <robertphillips@google.com>
>
TBR=bsalomon@google.com,robertphillips@google.com,reviews@skia.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
Change-Id: Iae6e313c15b2352bd0d4fc7b5629de0a51ac398e
Reviewed-on: https://skia-review.googlesource.com/9788
Reviewed-by: Yuqian Li <liyuqian@google.com>
Commit-Queue: Yuqian Li <liyuqian@google.com>
This unused feature complicates MDB.
Chrome compiles locally for me with this CL.
Change-Id: I611e464885fb984030eace43ead42cf39d0e7f72
Reviewed-on: https://skia-review.googlesource.com/9734
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
BUG=skia:6242
API change just makes already public API actually accessible.
TBR=reed@google.com
Change-Id: I72f9271db5f14e6fc35628bed5e16b01b304de4c
Reviewed-on: https://skia-review.googlesource.com/9743
Commit-Queue: Brian Osman <brianosman@google.com>
Reviewed-by: Brian Osman <brianosman@google.com>
Also disables use of small distance fields in Android framework.
Change-Id: I1ba40ce85aa34d067608587e1fbe1d42e8a42868
Reviewed-on: https://skia-review.googlesource.com/9731
Commit-Queue: Jim Van Verth <jvanverth@google.com>
Reviewed-by: Chris Dalton <csmartdalton@google.com>
This also makes it so that it doesn't shrink back into preallocated storage and therefore doesn't need to store the reserve count.
Change-Id: Ia320fed04c329641a5494947db39cefd2fb6d80f
Reviewed-on: https://skia-review.googlesource.com/9531
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Brian Salomon <bsalomon@google.com>
setMipColorMode seems like an odd call.
Change-Id: I24a1ac3883d52499f3be27282d006144d15b26f1
Reviewed-on: https://skia-review.googlesource.com/9725
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
This reverts commit eaaebb19a1.
Reason for revert: may call SkReader32::read(null, 0) -- reader needs to handle this
Original change's description:
> store vertices arrays inline with object
>
> Also unify some of naming (esp. around texCoords)
>
> BUG=skia:6366
>
> Change-Id: I5a6793f029cccf0cd0a2c1d180b259ce4eab526f
> Reviewed-on: https://skia-review.googlesource.com/9705
> Commit-Queue: Mike Reed <reed@google.com>
> Reviewed-by: Brian Salomon <bsalomon@google.com>
>
TBR=bsalomon@google.com,reed@google.com,reviews@skia.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:6366
Change-Id: Ie421654bcd74d74f8be6676291e3d6e16e2a7a16
Reviewed-on: https://skia-review.googlesource.com/9727
Reviewed-by: Mike Reed <reed@google.com>
Commit-Queue: Mike Reed <reed@google.com>
Also unify some of naming (esp. around texCoords)
BUG=skia:6366
Change-Id: I5a6793f029cccf0cd0a2c1d180b259ce4eab526f
Reviewed-on: https://skia-review.googlesource.com/9705
Commit-Queue: Mike Reed <reed@google.com>
Reviewed-by: Brian Salomon <bsalomon@google.com>
I did some testing with very large datasets, and the difference in max
RSS is measurable and significant.
Change-Id: I6bb2f795d5b4f6ebdba42c3089dc85a278355d48
Reviewed-on: https://skia-review.googlesource.com/9686
Commit-Queue: Hal Canary <halcanary@google.com>
Reviewed-by: Ben Wagner <bungeman@google.com>
and retract GrSurfaceContextPriv a bit
Change-Id: Id47af1052f9bda4fe7c85b3ce46b3ebe37797524
Reviewed-on: https://skia-review.googlesource.com/9647
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Robert Phillips <robertphillips@google.com>
precursor to enabling serialization in pictures
BUG=skia:6366
Change-Id: Iba89aa98f389b3281e7705d041e3337e89071f03
Reviewed-on: https://skia-review.googlesource.com/9680
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Mike Reed <reed@google.com>
Possible next iterations:
- remove another allocation use the SkData trick to share
the object and its (trailing) data
- store a bit that tells use to free each pointer, allowing
the builder to "adopt" some allocations instead of copy.
Larger idea:
- merge with drawPoints to have a single object for both.
BUG=skia:6366
Change-Id: Iec33239aa2ad5d00b36469ca0b88934ddf6f22eb
Reviewed-on: https://skia-review.googlesource.com/9604
Reviewed-by: Brian Salomon <bsalomon@google.com>
Commit-Queue: Mike Reed <reed@google.com>
